The dependent variables were computed as in Study 1.
Second, the concentration indices for these variables were computed to augment the concentration curves.
Thus, log transformation of variables (log 10) was done, and new variables were computed.
In addition to these variables, predictor variables were computed from the demographic data.
A total of 1455 variables were computed for each algebraic form (quadratic, bilinear and linear maps).
Then the partial (Type-3) R-square values for the independent variables were computed.
According to this data, the values for all the variables were computed.
The experimental results were analysed statistically and the main effects of the variables were computed and analysed.
To identify the dominant variables which most influence temperature response scatter, the correlation coefficients of various variables were computed.
For the two multivariable linear regression models, the total variance accounted by all predictor variables were computed.
The optimum coded values of three test variables were computed as x1 = 0.125, x2 = 0.707, and x3 = 0.107 by using a LOQO/AMPL optimization algorithm.
